PETERSON ET AL.: FOOD HABITS OF JUVENILE SALMON 



bers of fishes. Significantly greater numbers of 

 hyperiid amphipods occurred, however, in 

 larger than in smaller coho (Fig. 3). Average 

 weight of fishes in stomachs also increased with 

 length of juvenile coho. Total weight of stomach 

 contents was related to length of juvenile coho 

 (weight of prey = —1.0 + 0.016 X length of coho, 

 r = 0.43, P<0.01). The relationship between the 

 two food groups most important to juvenile coho 

 was investigated by plotting weights of euphau- 

 siids versus weights of fishes in the stomachs for 

 each of the 87 coho. This plot was divided into 

 quadrants by drawing lines parallel to the 

 abscissa and ordinate at the median values of 

 euphausiid and fish weight. The numbers of data 

 points in each quadrant are shown in a 2 X 2 con- 

 tingency table (Table 5). The x 2 of 12.5 was 

 highly significant (P<0.01, 1 df), indicating a 

 strong tendency for juvenile coho to contain large 

 amounts by weight of either fishes or euphausi- 

 ids, but not both at the same time. Of these 87 

 coho, 26% contained only fishes and 21% only 

 euphausiids. This trend may be a result of active 

 selection of one type of prey or a result of prey 



Table 5.— Contingency table comparing 

 weight of fish and euphausiid prey found in 

 the stomachs of 87 juvenile coho salmon. 



Euphausiid prey (g) 



<0.24 



-0.24 



Fish prey (g) 



>0.48 

 <0.48 



31 



13 



13 

 30 



patchiness. The latter explanation may be more 

 plausible, since stratification of these two prey 

 groups in the stomachs was evident in those indi- 

 viduals containing both prey items. 



Variability in the composition of stomach con- 

 tents of coho salmon was often high among the 10 

 stations where at least six fish were analyzed per 

 station (Table 6). For example, juvenile euphausi- 

 ids were the most numerically abundant prey 

 taxa at four stations (2, 3, 12, and 29); hyperiid 

 amphipods at four other stations (10, 27, 28, and 

 39); and fishes and crab larvae at one station 

 each. The differences in feeding habits among 

 stations were compared by calculating similarity 

 indices for all possible station pairs. We used the 

 percent similarity index (PSI = 1 min P { ), where 



ioo 



JUVENILE 

 EUPHAUSIIDS 



HYPERIIDS 



" I i ' "" I i 1 r 



IOO 140 180 220 260 300 



FISHES 



-f™p^l | H I I I ! 



IOO 140 180 220 260 300 



i r 



220 260 300 



SIZE CLASS OF SALMONID (mm) 



FIGURE 3.— Average abundance of juvenile euphausiids and hyperiids and average weight of fish prey occurring in the stomachs of 

 each 20 mm size class of juvenile coho and chinook salmon. The averages are taken over only those stomachs in which prey items 

 occurred. The numbers at the top of the leftmost figure denote the sample size in each 20 mm size class. 



845 



